fix(client): adding get states decode call, status command to check local gpud "/states", add sub-command aliases

client/v1/examples/get-states/main.go

@@ -0,0 +1,35 @@
+package main
+
+import (
+	"context"
+	"errors"
+	"time"
+
+	client_v1 "github.com/leptonai/gpud/client/v1"
+	"github.com/leptonai/gpud/errdefs"
+	"github.com/leptonai/gpud/log"
+)
+
+func main() {
+	baseURL := "https://localhost:15132"
+	for _, componentName := range []string{"disk", "accelerator-nvidia-info"} {
+		ctx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
+		defer cancel()
+		states, err := client_v1.GetStates(ctx, baseURL, client_v1.WithComponent(componentName))
+		if err != nil {
+			if errors.Is(err, errdefs.ErrNotFound) {
+				log.Logger.Warnw("component not found", "component", componentName)
+				return
+			}
+
+			log.Logger.Error("error fetching component info", "error", err)
+			return
+		}
+
+		for _, ss := range states {
+			for _, s := range ss.States {
+				log.Logger.Infof("state: %q, healthy: %v, extra info: %q\n", s.Name, s.Healthy, s.ExtraInfo)
+			}
+		}
+	}
+}

client/v1/v1.go

@@ -242,7 +242,7 @@ func GetStates(ctx context.Context, addr string, opts ...OpOption) (v1.LeptonSta
 		return nil, errors.New("server not ready, response not 200")
 	}
 
-	return nil, nil
+	return ReadStates(resp.Body, opts...)
 }
 
 func ReadStates(rd io.Reader, opts ...OpOption) (v1.LeptonStates, error) {

cmd/gpud/command/command.go

@@ -448,7 +448,7 @@ sudo rm /etc/systemd/system/gpud.service
 		},
 		{
 			Name:    "logs",
-			Aliases: []string{"l"},
+			Aliases: []string{"log", "l"},
 
 			Usage:  "checks the gpud logs",
 			Action: cmdLogs,
@@ -499,7 +499,7 @@ cat summary.txt
 		},
 		{
 			Name:    "scan",
-			Aliases: []string{"s"},
+			Aliases: []string{"check", "s"},
 
 			Usage:  "quick scans the host for any major issues",
 			Action: cmdScan,

cmd/gpud/command/run.go

@@ -121,7 +121,7 @@ func cmdRun(cliContext *cli.Context) error {
 
 	if pkd_systemd.SystemctlExists() {
 		if err := notifyReady(rootCtx); err != nil {
-			log.Logger.Warn("notify ready failed")
+			log.Logger.Warnw("notify ready failed")
 		}
 	} else {
 		log.Logger.Debugw("skipped sd notify as systemd is not available")

cmd/gpud/command/status.go

@@ -4,13 +4,16 @@ import (
 	"context"
 	"crypto/tls"
 	"encoding/json"
+	"errors"
 	"fmt"
 	"io"
 	"net/http"
 	"time"
 
 	client "github.com/leptonai/gpud/client/v1"
 	"github.com/leptonai/gpud/config"
+	"github.com/leptonai/gpud/errdefs"
+	"github.com/leptonai/gpud/log"
 	"github.com/leptonai/gpud/manager/packages"
 	"github.com/leptonai/gpud/pkg/systemd"
 
@@ -34,6 +37,16 @@ func cmdStatus(cliContext *cli.Context) error {
 	}
 	fmt.Printf("%s successfully checked gpud status\n", checkMark)
 
+	if err := checkDiskComponent(); err != nil {
+		return err
+	}
+	fmt.Printf("%s successfully checked whether disk component is running\n", checkMark)
+
+	if err := checkNvidiaInfoComponent(); err != nil {
+		return err
+	}
+	fmt.Printf("%s successfully checked whether accelerator-nvidia-info component is running\n", checkMark)
+
 	if err := client.BlockUntilServerReady(
 		rootCtx,
 		fmt.Sprintf("https://localhost:%d", config.DefaultGPUdPort),
@@ -76,6 +89,59 @@ func cmdStatus(cliContext *cli.Context) error {
 	return nil
 }
 
+func checkDiskComponent() error {
+	baseURL := fmt.Sprintf("https://localhost:%d", config.DefaultGPUdPort)
+	componentName := "disk"
+
+	ctx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
+	defer cancel()
+	states, err := client.GetStates(ctx, baseURL, client.WithComponent(componentName))
+	if err != nil {
+		// assume disk component is enabled for all platforms
+		return err
+	}
+	if len(states) == 0 {
+		log.Logger.Warnw("empty state returned", "component", componentName)
+		return errors.New("empty state returned")
+	}
+
+	for _, ss := range states {
+		for _, s := range ss.States {
+			log.Logger.Infof("state: %q, healthy: %v, extra info: %q\n", s.Name, s.Healthy, s.ExtraInfo)
+		}
+	}
+
+	return nil
+}
+
+func checkNvidiaInfoComponent() error {
+	baseURL := fmt.Sprintf("https://localhost:%d", config.DefaultGPUdPort)
+	componentName := "accelerator-nvidia-info"
+
+	ctx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
+	defer cancel()
+	states, err := client.GetStates(ctx, baseURL, client.WithComponent(componentName))
+	if err != nil {
+		if errors.Is(err, errdefs.ErrNotFound) {
+			log.Logger.Warnw("component not found", "component", componentName)
+			return nil
+		}
+		return err
+	}
+	if len(states) == 0 {
+		log.Logger.Warnw("empty state returned", "component", componentName)
+		return errors.New("empty state returned")
+	}
+
+	for _, ss := range states {
+		for _, s := range ss.States {
+			log.Logger.Infof("state: %q, healthy: %v, extra info: %q\n", s.Name, s.Healthy, s.ExtraInfo)
+		}
+	}
+
+	return nil
+}
+
 func getStatus() ([]packages.PackageStatus, error) {
 	httpClient := &http.Client{
 		Transport: &http.Transport{

e2e/e2e_test.go

@@ -283,6 +283,17 @@ func TestGpudHealthzInfo(t *testing.T) {
 	}
 	t.Logf("respMetrics size:\n%s", string(metricsBytes))
 
+	t.Log("now testing with client/v1 basic disk get states")
+	states, err := client_v1.GetStates(ctx, "https://"+ep, client_v1.WithComponent("disk"))
+	if err != nil {
+		t.Errorf("failed to get states: %v", err)
+	}
+	for _, ss := range states {
+		for _, s := range ss.States {
+			t.Logf("state: %q, healthy: %v, extra info: %q\n", s.Name, s.Healthy, s.ExtraInfo)
+		}
+	}
+
 	t.Log("now testing with client/v1")
 	for _, opts := range [][]client_v1.OpOption{
 		{client_v1.WithRequestContentTypeJSON()},